llvm-svn: 134298

llvm-svn: 134287

up the valid constant check earlier.

rdar://9692967

llvm-svn: 134286

llvm-svn: 134282

llvm-svn: 134281

llvm-svn: 134279

outside the loop and reducible.

This more completely hides them from LSR, which isn't usually able to
do anything meaningful with non-affine expressions anyway, and this
consequently hides them from SCEVExpander, which is acutely unprepared
for non-affine expressions.

Replace test/CodeGen/X86/lsr-nonaffine.ll with a new test that tests
the new behavior.

This works around the bug in PR10117 / rdar://problem/9633149, and is
generally an improvement besides.

llvm-svn: 134268

Generalize @llvm.ctlz, @llvm.cttz, and @llvm.ctpop to work on vectors of integers, and fix the one optimization pass that I'm aware of that needs updating for this.  At least one current target, ARM NEON, can implement these operations on vectors directly.

llvm-svn: 134265

llvm-svn: 134264

The DSP instructions in the Thumb2 instruction set are an optional extension
in the Cortex-M* archtitecture. When present, the implementation is considered
an "ARMv7E-M implementation," and when not, an "ARMv7-M implementation."

Add a subtarget feature hook for the v7e-m instructions and hook it up. The
cortex-m3 cpu is an example of a v7m implementation, while the cortex-m4 is
a v7e-m implementation.

rdar://9572992

llvm-svn: 134261

llvm-svn: 134259

  itineraries.
- Refactor TargetSubtarget to be based on MCSubtargetInfo.
- Change tablegen generated subtarget info to initialize MCSubtargetInfo
  and hide more details from targets.

llvm-svn: 134257

(low two bits always zero, so off by one bit of encoded value).

llvm-svn: 134247

llvm-svn: 134244

t2MOVCC[ri] are just t2MOV[ri] instructions, so properly pseudo-ize them.
The Thumb1 versions, tMOVCC[ri] were only present for use by the size-
reduction pass, so they're no longer necessary at all and can be deleted.

llvm-svn: 134242

llvm-svn: 134240

copy is a kill") to see if it fixes the i386 dragonegg buildbot, which is timing out
because gcc built with dragonegg is going into an infinite loop.

llvm-svn: 134237

llvm-svn: 134236

llvm-svn: 134235

llvm-svn: 134234

llvm-svn: 134231

llvm-svn: 134229

llvm-svn: 134228

The constraints are represented by the register class of the original
virtual register created for the inline asm. If the register class were
included in the operand descriptor, we might be able to do this.

For now, just give up on regclass inflation when inline asm is involved.

No test case, this bug hasn't happened yet.

llvm-svn: 134226

llvm-svn: 134224

llvm-svn: 134223

supporting the instruction that the constraint is for 'movw'.

Part of rdar://9119939

llvm-svn: 134222

llvm-svn: 134221

for the 'x' register constraint.

Part of rdar://9119939

llvm-svn: 134220

llvm-svn: 134219

Part of rdar://9119939

llvm-svn: 134217

llvm-svn: 134216

Part of rdar://9307836 and rdar://9119939

llvm-svn: 134215

llvm-svn: 134212

llvm-svn: 134211

llvm-svn: 134210

llvm-svn: 134209

llvm-svn: 134208

We would put the return value from long double functions in the wrong
register.

This fixes gcc.c-torture/execute/conversion.c

llvm-svn: 134205

Merge the tMOVr, tMOVgpr2tgpr, tMOVtgpr2gpr, and tMOVgpr2gpr instructions
into tMOVr. There's no need to keep them separate. Giving the tMOVr
instruction the proper GPR register class for its operands is sufficient
to give the register allocator enough information to do the right thing
directly.

llvm-svn: 134204